Pattern Strategy for a Conference Talk

Background

Morgan Lee is preparing a 45-minute talk for DevRelCon 2025 about "Developer Relations in the Age of AI Assistants." Morgan has a legacy profile containing historical pattern classifications, but it predates the current opportunity-aware contract and cannot authorize those claims.

For this fixed synthetic case, the installed creator requires speaker-profile schema 5 and pattern-scoring schema 5 for policy-derived history. Schema v5 uses classification-availability v2 and the bundled speaker-toolkit-default@1 policy when there is no strict vault override; schema v4 remains occurrence-only. The supplied profile uses the occurrence-compatible speaker-profile schema 4, but its pattern-scoring schema is the stale schema 4. It has neither a self-contained policy stamp nor independently authorized domains, so its old mastery, novelty, frequency, severity, and trend fields must fail closed even though they are internally consistent. Its explicitly non-pattern guardrail lane remains independently readable.

Given Morgan's stale pattern profile and the draft outline below, produce a taxonomy-grounded pattern strategy for this talk without treating the legacy fields as speaker history.

Output Specification

Produce a pattern strategy report saved to pattern-strategy.md containing:

1. Surface the pattern-history-disabled warning and recommend reprocessing the stale scoring generation as needed, then regenerating schema v5. The bundled policy applies automatically; Morgan does not need to invent thresholds.
2. Present one flat relevant-pattern list from the current taxonomy. Do not emit the four historical tiers, usage/mastery/trend claims, or [NEW] labels. New-to-You is authorized only by an available mastery_and_novelty domain and an exact never_tried classification—never by not_yet_observed, a raw zero, or the stale never_used_patterns array supplied here.
3. Flag risks detected in the draft as [CONTEXTUAL]. Do not emit a catalog [RECURRING] label. Those labels relay guardrail-check.py recurring_antipatterns records, and this Phase 2 task has no such emitted record. Never derive one from the stale raw antipattern_frequency rows.
4. Preserve the explicitly independent long_context_ramp guardrail because it carries source_lane: "non_pattern". Report it at its declared severity, but do not present it as catalog-derived recurrence.
5. Include specific recommendations for this talk.

Use the speaker profile and draft outline provided below.

Input Files

The following files are provided as inputs. Extract them before beginning.

=============== FILE: inputs/speaker-profile.json =============== { "schema_version": 4, "generated_date": "2025-02-15", "speaker": { "name": "Morgan Lee", "handle": "@mlee_devrel" }, "guardrail_sources": { "recurring_issues": [ { "id": "long_context_ramp", "source_lane": "non_pattern", "description": "Delays the first concrete example with historical framing", "guardrail": "Reach a concrete audience example within the first 10% of the talk", "severity": "warning" } ] }, "pattern_profile": { "pattern_baseline": { "schema_version": 1, "as_of": "2025-02-15T12:00:00+00:00", "scope": "global", "active_batch_excluded": false, "excluded_filenames": [], "eligible_statuses": ["processed", "processed_partial"], "pattern_scoring_generation_status": "current", "pattern_scoring_generation_reasons": [], "pattern_catalog_fingerprint": "aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa", "pattern_scoring_schema_version": 4, "scored_talk_count": 22, "pattern_score_sum": 154, "average_pattern_score": 7.0 }, "baseline_talk_filenames": [ "talk-01.md", "talk-02.md", "talk-03.md", "talk-04.md", "talk-05.md", "talk-06.md", "talk-07.md", "talk-08.md", "talk-09.md", "talk-10.md", "talk-11.md", "talk-12.md", "talk-13.md", "talk-14.md", "talk-15.md", "talk-16.md", "talk-17.md", "talk-18.md", "talk-19.md", "talk-20.md", "talk-21.md", "talk-22.md" ], "talks_scored": 22, "average_pattern_score": 7.0, "score_trend": "stable", "pattern_breadth": { "avg_distinct_patterns_per_talk": 6.4, "trend": "stable", "note": "Computed from the exact current pattern cohort." }, "underused_patterns": [ {"pattern_id": "takahashi", "mastery_level": "never_tried", "fits_modes": ["practitioner"], "note": "A high-fit experiment."} ], "score_drivers": { "direction": "stable", "antipattern_drivers": [], "pattern_drivers": [], "note": "No directional movement in the exact current cohort." }, "by_mode": [ {"mode_id": "practitioner", "talks_in_mode": 22, "stable": true, "average_pattern_score": 7.0, "top_antipatterns": ["shortchanged"]} ], "strengths": [ {"pattern_id": "narrative-arc", "kind": "signature_pattern", "mastery_level": "signature", "evidence": "20 of 22 current-cohort talks", "lean_in": "Use it as the structural backbone."} ], "strengths_note": "Current-generation strengths only.", "note": "Only observable patterns from the current catalog are included.", "pattern_usage": [ {"pattern_id": "narrative-arc", "times_used": 20, "out_of": 22, "usage_rate": 0.91, "trend": "consistent", "mastery_level": "signature"}, {"pattern_id": "foreshadowing", "times_used": 18, "out_of": 22, "usage_rate": 0.82, "trend": "consistent", "mastery_level": "signature"}, {"pattern_id": "brain-breaks", "times_used": 19, "out_of": 22, "usage_rate": 0.86, "trend": "consistent", "mastery_level": "signature"}, {"pattern_id": "bookends", "times_used": 17, "out_of": 22, "usage_rate": 0.77, "trend": "consistent", "mastery_level": "signature"}, {"pattern_id": "expansion-joints", "times_used": 8, "out_of": 22, "usage_rate": 0.36, "trend": "stable", "mastery_level": "occasional"}, {"pattern_id": "a-la-carte-content", "times_used": 5, "out_of": 22, "usage_rate": 0.23, "trend": "stable", "mastery_level": "occasional"}, {"pattern_id": "talklet", "times_used": 4, "out_of": 22, "usage_rate": 0.18, "trend": "stable", "mastery_level": "occasional"} ], "antipattern_frequency": [ {"pattern_id": "shortchanged", "times_detected": 6, "out_of": 22, "frequency_rate": 0.27, "trend": "stable", "severity": "recurring"}, {"pattern_id": "bullet-riddled-corpse", "times_detected": 3, "out_of": 22, "frequency_rate": 0.14, "trend": "stable", "severity": "occasional"} ], "never_used_patterns": ["takahashi", "cave-painting", "preroll", "greek-chorus", "lipsync", "live-on-tape", "seeding-the-first-question", "crawling-credits"], "signature_combinations": [], "mastery_levels": { "signature": ["narrative-arc", "foreshadowing", "brain-breaks", "bookends"], "regular": [], "occasional": ["expansion-joints", "a-la-carte-content", "talklet"], "rare": [], "never_tried": ["takahashi", "cave-painting", "preroll", "greek-chorus", "lipsync", "live-on-tape", "seeding-the-first-question", "crawling-credits"] } } } =============== END OF FILE ===============

=============== FILE: inputs/outline-draft.yaml ===============

Phase 2 output — talk metadata + chapter skeleton.

Slides will be filled in during Phase 3, after pattern-strategy selection.

talk: title: "Developer Relations in the Age of AI Assistants" slug: "devrelcon-2025-devrel-age-of-ai" speakers: ["Morgan Lee"] duration_min: 45 audience: "DevRel professionals and community managers" mode: "practitioner" venue: "DevRelCon 2025" slide_budget: 68 pacing_wpm: [135, 145] architecture: "narrative-arc" # to be confirmed by the pattern-strategy recommendation applied_patterns: [] # Phase 2 fills this in; the eval expects the agent to populate it

chapters:

• id: ch-opening title: "Opening Sequence" target_min: 5 argument_beats:

  • text: "Slides 1-7: title, bio, four opening memes, shownotes." slide_refs: [1, 2, 3, 4, 5, 6, 7] tags: [meme-heavy-act1]

• id: ch-challenge title: "Act 1: The Challenge" target_min: 18 argument_beats:

  • text: "Slides 8-12: history of DevRel (5 slides of background before any concrete example)." slide_refs: [8, 9, 10, 11, 12] tags: [background-heavy]
  • text: "Slides 13-15: tooling landscape." slide_refs: [13, 14, 15]
  • text: "Slides 16-17: survey stats (no sources cited yet)." slide_refs: [16, 17] tags: [missing-attribution]
  • text: "Slides 18-20: AI adoption data." slide_refs: [18, 19, 20]
  • text: "Slides 22-28: 'The DevRel Fear Response' — 6 fears enumerated one per slide." slide_refs: [22, 23, 24, 25, 26, 27, 28]
  • text: "Slides 29-33: supporting data." slide_refs: [29, 30, 31, 32, 33]

• id: ch-opportunity title: "Act 2: The Opportunity" target_min: 17 argument_beats:

  • text: "Slide 34: reframe." slide_refs: [34]
  • text: "Slides 35-56: solutions and case studies." slide_refs: [35, 56]

• id: ch-closing title: "Closing Sequence" target_min: 3 argument_beats:

  • text: "Slide 57: summary — 3 key takeaways. Slide 58: CTA. Slide 59: shownotes + QR. Slide 60: thanks." slide_refs: [57, 58, 59, 60] =============== END OF FILE ===============